Skip to main content

Python package for ocean acoustics modelling

Project description

Welcome to Kadlu, a Python package for modelling ocean acoustics

Kadlu was developed for the purpose of modelling noise due to waves and rain in shallow coastal waters, but contains tools useful for many other underwater acoustics modelling tasks.

Kadlu is written in Python and utilizes a number of powerful software packages including NumPy, HDF5, NetCDF-4, and SQLite. It is licensed under the GNU GPLv3 license and hence freely available for anyone to use and modify. The project is hosted on GitHub at github.com/meridian-acoustics/kadlu.

Kadlu was born out of the MERIDIAN project (2018-2023) at the Institute for Big Data Analytics at Dalhousie University with the support and assistance of David Barclay and Calder Robinson, both from the Department of Oceanography at Dalhousie University.

Kadlu provides functionalities that automate the process of fetching and interpolating environmental data necessary to model ocean ambient noise levels (bathymetry, water temperature and salinity, wave height, wind speed, etc.). It also includes various routines that allow accurate estimates of noise source levels and transmission losses in realistic ocean environments. You can find more information about the technical aspects of how sound propagation is modelled in Kadlu in this note.

The intended users of Kadlu are researchers and students in underwater acoustics working with ambient noise modeling. While Kadlu comes with complete documentation and comprehensive step-by-step tutorials, some familiarity with Python and especially the NumPy package would be beneficial. A basic understanding of the physical principles of underwater sound propagation would also be an advantage.

For more information, please consult Kadlu's Documentation Page.

Installation

Kadlu runs on the most recent stable version of Python 3.

  1. ensure that GCC is installed. On Ubuntu, this can be done by installing the build-essential package if its not installed already

    sudo apt update 
    sudo apt install build-essential
    
  2. Install Kadlu using pip

    python -m pip install kadlu
    

Configuration

Optionally set the storage directory

Kadlu allows configuration for where data is stored on your machine. By default, a folder 'kadlu_data' will be created in the home directory. To specify a custom location, run the following:

import kadlu
kadlu.storage_cfg(setdir='/specify/desired/path/here/')

Optionally add an API token for fetching ERA5 data

Kadlu uses ECMWF's ERA5 dataset as one its main sources of environmental data, e.g., for wind, waves and solar irradiance. You will need to obtain your own access token to download these data. This can be done by registering an account with ECMWF and the Copernicus Climate Data Store (CDS). Once logged in, your access token, comprised of an URL address and a secret key, will be displayed on the Copernicus webpage. Following the instructions, copy the URL and the key to the file $HOME/.cdsapirc. Finally, you will need to accept the ERA5 terms of use to activate the token.

Jupyter notebook tutorials

  1. The Ocean Module

  2. Fetch and Load Environmental Data

  3. Interpolate Multi-Dimensional Data

  4. Plot and Export Data

  5. Transmission Loss

Notes for developers

The CI pipeline caches all data pulled from external APIs by the tests. This allows for faster execution of the pipeline, but means that the data fetching step is not being tested. Therefore, it is advisable to periodically clean the runner caches manually.

To skip execution of the CI pipeline, append -o ci.skip to the git push command.

Steps for publishing a new release to PyPi:

python setup.py sdist bdist_wheel
twine upload dist/kadlu-X.Y.Z.tar.gz

Useful resources

  • gsw Python package (Python implementation of the Thermodynamic Equation of Seawater 2010)

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

kadlu-2.4.2.tar.gz (104.9 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page